INDEX

Explanations

fragments with the word "thought"

oai_token-act-pair · gpt-3.5-turbo

expressions of personal thoughts or opinions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

conservancy

-0.75

kw

-0.65

\-

-0.64

 clad

-0.62

 Peninsula

-0.62

versions

-0.62

acerb

-0.61

iona

-0.60

 legion

-0.60

clad

-0.60

POSITIVE LOGITS

fully

0.97

lessly

0.93

fulness

0.85

 provoking

0.77

rison

0.74

ileaks

0.72

ventus

0.68

culus

0.66

ful

0.66

ilyn

0.65

Activations Density 0.039%

fragments with the word "thought"

expressions of personal thoughts or opinions

No Comments

No Known Activations